Why 80% of UK Marketing Agencies Will Be Dead or Acquired by 2028

Spend twelve years inside the UK marketing agency industry and you learn to recognise the moments when a sector is about to be restructured before the people inside it do. The British high street had its moment in 2017. UK retail banking had its moment in 2019. The minicab industry had its moment in 2014, with Uber. The agency industry's moment is right now.

I am writing this from inside the disruption — Creative Agent, the agency I founded and rebuilt as AI-first in 2025, is twelve months into a transformation that has changed the economics so completely I no longer believe the legacy agency P&L is a defensible structure. The thesis below is uncomfortable for the industry I came from. I think it is also true.

1. The £19bn that is about to compress

The UK marketing services market is roughly £19bn in annual revenue, served by approximately 26,000 agencies. The shape of the market is bottom-heavy: a small handful of holding-company giants, a thin layer of mid-tier independents (50–250 staff), and a long tail of small shops under twenty people. The mid-tier is where most of the money is made and where most of the disruption lands.

2. Force one: AI cost curves cut deliverable cost by 90%

The cost to produce a 1,500-word SEO article in 2022 was a writer's day plus an editor's hour — call it £400. The cost to produce one in 2026, with the right agentic workflow, is roughly £40 of compute and ten minutes of senior review. The same compression has now reached creative production, paid-media account management, reporting, lifecycle email, and competitive research.

That is not "AI helps the team work faster". That is the deliverable cost falling by an order of magnitude. The agency that keeps charging 2022 prices in a 2026 cost market gets undercut by the challenger that prices on the new cost base. The agency that keeps the same headcount on the same workload runs out of margin. The economic squeeze is not optional.

3. Force two: in-housing has gone from "considered" to "default"

The 2024 IPA Bellwether already showed in-housing accelerating. ChatGPT in the marketing director's pocket has accelerated it again. Brands that would have outsourced content in 2023 now generate it internally. Brands that would have hired an SEO agency in 2024 now have a single in-house lead operating an AI stack. Brands that would have run their paid media through an agency in 2025 now use Performance Max + a junior + a senior sign-off.

Total agency demand falls — not catastrophically, but structurally — by an estimated 25–35% by 2028. That is the margin the average mid-tier agency was running on. It does not come back.

4. Force three: the time-and-materials retainer breaks

The third force is the most under-appreciated. The retainer model — £15k a month for "everything", priced against an FTE assumption — was always an awkward fiction. AI breaks the fiction completely. When the client knows the deliverable costs £40 of compute, paying £15k for it requires a story the agency can no longer tell with a straight face.

The retainer survives only by repricing onto outcomes. I wrote a separate confession on this from inside the agency P&L — the short version is that retainer revenue moves to results-based pricing or it disappears.

5. The compounded survival rate

Each force on its own is survivable. Together they compound. An agency that loses 30% of demand to in-housing, sees its deliverable cost base undercut by AI-native challengers, and has its core retainer line repriced — all simultaneously — has only two viable moves: restructure the P&L onto outcomes, or sell.

Most agency owners will choose neither. They will hope. The hope strategy works for about 18 months. After that, the cashflow forces the decision.

6. The three archetypes that survive

• Outcome-priced challengers. Agencies whose P&L is already built on results. Creative Agent is one. They absorb the in-housing wave because they were never selling time. They absorb the AI cost curve because they were never passing it on.
• Deep specialists. Agencies in narrow verticals — UK pharma marketing, regulated financial services, B2B defence — where domain depth and trust create a moat AI does not yet cross. These agencies stay roughly the same size and get more profitable.
• Consolidated rollups. Buyers who aggregate smaller agencies, rationalise cost aggressively, and run them on a shared AI-native delivery layer. Creative Marketing's acquisitions arm is built for this moment.

7. Why this is the best UK agency M&A window in a decade

Seller expectations are still anchored to 2021–2022 valuations, when a healthy mid-tier agency traded at 6–8x EBITDA and an ambitious one at 10x. Buyer economics in 2026 are based on a cost base that is 90% lower for the production layer. The arbitrage is enormous and it closes by 2027 as seller expectations adjust to the new market.

I expect the smartest UK agency rollups in history to happen in the next 18 months. After that, the buyers and sellers will be negotiating from the same information set, and the deals get ordinary again.

8. What you should do, depending on who you are

If you own a UK agency — model your P&L under all three forces simultaneously, not one at a time. Decide, honestly, whether your business is one of the three surviving archetypes. If it is not, the right move in 2026 is to sell into a rollup while the multiples are still arbitraged. The Creative Marketing acquisitions team looks at every UK agency that sends us a deck.

If you run a marketing function inside a brand — the agency you used in 2024 may not exist in 2027. Build optionality. Test outcome-priced challengers on a small portion of the budget now, before you need them.

If you are a journalist or analyst — the story is not "AI inside agencies". The story is the structural collapse of the seat-priced agency P&L and the consolidation window that opens behind it.

To talk through any of this, book a call. To go deeper on the agency-economics thesis, read The Death of the £15k/Month Retainer and The Future Agency Has 12 People and 200 Agents.